CMM Calibration with PCDMIS

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • CMM Calibration with PCDMIS

    What exactly is involved when a CMM is calibrated? What do they calibrate? Also how will that affect PCDMIS? We are having ours calibrated and would like to know ahead of time of any possible issues that can arise with PCDMIS.
    I used to be high on life but I built up a tolerance.

    Brown & Sharpe Global Advantage
    PCDMIS CAD++ v2011mr2
    PH10MQ/SP600M


    sigpic

  • #2
    Who is doing the calibration and how, what type? Laser? Ball-bar? Hammer and tape measure?
    Without knowing this, it is kind of hard to give an answer.

    Also, what controller?
    sigpic
    Originally posted by AndersI
    I've got one from September 2006 (bug ticket) which has finally been fixed in 2013.

    Comment


    • #3
      Also what type of CMM? is it a Brown and Sharp? or another machine brand that has a different controller? Who is doing the calibration B&S or just a local calibration house that says they can calibrate it but what they really mean is that they can verify that it is still within manufactures spec's.


      BS
      Windows 7
      Pc-Dmis 2015
      Global Performance 7-10-7


      Comment


      • #4
        There are also different standards that people calibrate cmm's to (i.e. the B89 standard, VDI/VDE etc...) so be aware of what they calibrate to. I have also found that sometimes they will short-change you as well.

        A brief summary of what could be done(the order may vary) is to first clean the machine(bearing ways, the scales as well. Next check mechanical squareness of each axis to the other. Some will also make sure the signal to the reader head is good and will adjust(align the reader head) to optimize the signal. A repeatability on the sphere is also done. Measure linearity of each axis. Some will use a ballbar - some use gage blocks.

        Overall one suggestion that I would say is to have a "master part" - meaning a part that has been measured and verified by another means if possible. This part should be checked right before they start calibration. When they are done have them get a coffee and then check your part. If it measures fine then you know things are OK. If not then I would be cautious of what changed and be making sure they are not out the door on you and now you have a machine that could be in question.

        Those are my .02 and I have seen people get burned by the wannabe third parties. In fact one time I was told a third party came in and measured a 10" gage block for all 3 axis and the sphere and made up some report and called it certified!!!! Be careful........

        Comment


        • #5
          The more you know about metrology the better equipped you will be to know if your calibration is really a calibration.

          Your machine ways need to be straight and orthogonal. You need a steady signal to your reader heads and you need repeatability.

          A good calibration will check for geometry errors, repeatability and accuracy.

          More than a few third party calibration houses really do not have the expertise to calibrate your machine. Many of them can tell you something is wrong but a lot of them cannot fix the problem.

          We use Brown and Sharpe to calibrate our machines and the techs I have come across do know what they are doing......they just need to know that YOU know what they are doing.

          Gageguru has a good idea and one that I use in that I have a sample part that I measure periodically both over the short term ( in a loop ) and over the long term (weekly ) and I keep two sets of control charts so I can see when my machine is changing.

          There are a couple of things that are important and you should not listen when some third party calibrator tells you differently.

          1) Accuracy-Determine accuracy with a laser or precision step gage that is calibrated by NIST or other prime standard lab.

          2) Orthogonality- Check with ball bar or precision square and electronic indicator.

          3) Volumetric accuracy- Ball bar........IF the actual length between ball centers is KNOWN. If this value is not known, you just repeated number 2.

          Calibration may be REQUIRED if the machine is bumped or crashed or if a reader head goes out and is replaced. I have not been convinced that one can change a reader head and reset the air gap and be certain the calibration has not been changed.

          Remember the machine can be very repeatable and WRONG or it can be repeatable and known within a documented uncertainty.

          You must know what to look for or you must know the calibration source business practice. Generally you will get what you pay for.

          Hope this helps.

          Hilton
          Hilton Roberts

          "Carpe Cerveza"

          Comment


          • #6
            Have the tech print out the results of all the measurements. I had a guy here and he could not even calibrate a probe and this guy was supposely a B&S employee. To make a long story short I caught the guy tweaking the results for the B89 Ball Bar. Needless to say I had a few choice words with him and Malcom Parson (in charge of customer service in my area). Now B&S is under a mircoscope with me for calibration. I only will allow a few tech's in our facility from B&S (ones that I know or refered to me by the sales rep.) But I still require a print out for each measurement from PC-DIMIS so I can compare each of the measurement, to the report they give you. Also, I am watching over their shoulder the whole time they are here and I will continue to do that from now on.
            Global Status 12-22-10
            Global Performance 9-15-8
            CAD++ 2015.1_SP7

            Comment


            • #7
              You hit on an interesting one. Calibration is NOT the same as re-comping your machine! However, it seems that all vendors make them the same.

              Fundamentally, calibration means that someone tests the accuracy of your machine against a standard. There are many standards , one of the most popular (and in my opinion) best, is B89 ball bar.

              What happens is that the machine changes shape over time. Just like me, it sags and becomes less pleasing to look at. Bolt joints relax slightly, materials relieve themselves slightly (and I mean stress relieve for all you smarties out there), air bearings get slightly clogged etc etc. So after some time it is reasonable to expect your machine has moved a little. So over time the performance against a standard will get worse. So you need a nip and tuck. Which is fundamentally what a re-comp is. In the old days, when I was young and beautiful, they unbolted joints and readjusted mechanically until all was perfect. Now, they can do that through software, which is better because all bolt joints relax over time. It is better to never touch them because after a while they become quite stable.

              So when they calibrate your machine, they will give you performance against a standard. Problem is, what if the machine does not meet the published requirements (or as they say: the machine is out)? That's when they go do a re-comp. Followed by a new calibration.

              Be warned about some real bad practices out there! I've been there and probable done them all. There are people out there that run a ball bar and find that 1 position is out. They change a squarenss comp number and state that your machine is back in. THAT IS TOTAL BALONY! If they change 1 comp number, you MUST DEMAND that they re-do the complete calibration all over again. Insist that during a calibration run, NONE of the comp values can be adjusted. Also, make sure they do not play with the temp comp numbers. I found one vendor that I specified him to do 20 ball bar positions. When I looked at the results, he told me that some of the linears had been out, but the machines temperature must have been changed. So he adjust the temp comp, so they were in. One he ran at 19 degrees C, the other at 22! My room was 20 +/- .25 at that time. So after that BS, I made him re-comp the whole machine; he had to relaser it. Recomping was part of the contract, so I made him do it.

              There are some real good people out there, and some crooks. Be very careful with the lowest cost guys.

              What I would do is this:
              Depending on the machine size, have him run at least 20 ( for a 765 machine) or at least 28 ball bar positions (for larger ones). Be aware that YOU as the user can specify the number of positions and where he must take them! Do not fall for their trap that says that they specifiy how many and where. That is again baloney. You specify the where and how many. Have them run a calibration and give you the number. If the machine is in, make him go away. DO NOT HAVE HIM TOUCH ANY OF YOUR COMPS IN THAT CASE (sometimes he will want to make it slightly better). If you fail, have him re-comp, and run a new calibration. Make sure you are there when he runs the new calibration. Because his management will quickly push him to futs around with comps to get out of there! And it is very difficult for you to find that.

              But also be realistic. If all your ball bars are in, in the area where you always use the machine, if you have a few slightly out at the extremes, who really cares? You probably have a good machine. So be realistic. Do not have him re-comp if some extreme case does not meet spec. And that then becomes your decision. Not the vendor's.

              I guess, what I want to tell you is that this process is something you need to be actively involved with, if you want it to be done right. Do not leave the guy alone! There are too many pressures on him to fool you. And some of them will do...


              Jan.
              ***************************
              PC-DMIS/NC 2010MR3; 15 December 2010; running on 18 machine tools.
              Romer Infinite; PC-DMIS 2010 MR3; 15 December 2010.

              Comment


              • #8
                Jan,

                You are a kindred spirit if I have ever run across one. You are absolutely right when you state you really do not want to turn any wrenches unless you are prepared to make the machine a whole lot worse before you can make it better.

                I remember in the middle '70s working on some Shelton CMMS that had big elevated granite rails for the X axis. They were out of straight and we would twist and pull them and finally get them in spec and go away for an hour and the machine would change.

                Those kinds of adjustments are ideal for electronic tweaking. Once the machine settles in, it will require only minor tweaking if any at all.

                All this baloney about adjusting "closer to nomnal" needs to be removed from the lexicon for calibration. More CMMS and hand tools have been damaged by excessive tweaking than were ever damaged by overuse.

                All one has to do is follow ones own work. If you are in the calibration business and get the chance to see your own work come back in for recertification, a little perusal of the calibration records will generally show that calibration techs do less adjusting when they follow their own work than they do when they follow the work of another tech.

                You seem to been around metrology for a while as your writing reflects much of what I have seen over the years.....and I have seen a lot.

                What was especially interesting to me was the last two machines we bought.

                After we calibrated them in place here in Mesa, we did a full up calibration after they had been here a year. The machines had changed a little bit and they were adjusted. Other than minor comp changes, the machines have been good since. We monitor the performance and keep careful track of the control charts. Once the machines settled in, they perform well and they are consistent.

                Hope to read many more of your posts!

                Hilton
                Hilton Roberts

                "Carpe Cerveza"

                Comment


                • #9
                  Oh yes, the Sheltons. I ran into several over the years. I worked on one at Boeing. I loved those machines. They were huge and they had all these obvious and very intellegent mechanical solution to make machines mechanically as good as they could get them. Russ was smart. Somehow, I always felt I was working on a steam engine or something: no sense of styling, just pure guts hanging on fuctional parts. Ah, those were the days...

                  I never had to straighten a rail, thank goodness. But I mechanically adjusted sqaureness many times. Typically, you did that by adjusting the airbearings. So you had it square, and then you had to tighten the lock screw! And the whole thing was out of sqaure again AAAAAAAAHHHHH..... Lost a lot of hair in those days.

                  We have the same experience as you have with those 2 new machines. Here, a new machine gets calibrated after run-off, and after 3 months. After that, once a year. Seems to work. They remain fairly stable.



                  Jan.
                  ***************************
                  PC-DMIS/NC 2010MR3; 15 December 2010; running on 18 machine tools.
                  Romer Infinite; PC-DMIS 2010 MR3; 15 December 2010.

                  Comment


                  • #10
                    Jan,

                    Russel Shelton had a lab in Paducah, Kentucky ( where I grew up ). Tried to go to work for them once. Didn't work out. McDonnell-Douglas it seems bought 3 prototype machines. First one had a round Z axis ram. Z axis movement was terrible. Had these little nets over the top of the rails. Put that thing in the shop with all the machine tools and then put a titanium deburr area right across the aisle. Titanium dust and coolant mist blew into the air and settled......on top of the X axis rails. The shop would not wipe down the machine so the titanium dust rolled up with the coolant mist keeping it all together and ground up the tops of the ways and ruined the machine.

                    We put another machine into the tooling area next to an outside wall and railroad tracks. This one had laser scales. Now this was slick...real slick The laser shot down X axis and went into a splitter that pulled off about 2/3rds of the signal and routed that along Y axis and then split half of that signal off for Z axis. You will love this part......I know I did....
                    Put a stainless steel corner reflector into an aluminum mount and fasten it with steel screws to a granite rail. After some time...usually about a day, the temperature extremes and the differences in coefficients of expansion caused the relectors to lose the beam. If you lost the Z axis reflector, you lost Z axis. If you lost the Y axis reflector, you lost Y and Z and BY GOLLY if you lost X axis reflector.....'ya lost the whole enchilada. Everytime you had to adjust Z square to X or Y, you would lose at LEAST one axis of motion. Usually more.

                    I do not know if Russ is still alive or not. I know the building is still there in Paducah but it is not called Shelton Metrology anymore. The last I heard of Russel, he was working for Zeiss or maybe Leitz. I am not sure. What I do know is that he had a vision and we are all benefitting from that. It was really a struggle at the time because we just got into too big a rush to hang lasers on everything. He was ahead of his time.

                    Hilton
                    Hilton Roberts

                    "Carpe Cerveza"

                    Comment


                    • #11
                      Wow, great story, thanks Hilton. I never met Russ, but I heard a great deal about him. He sure was well known in the 80s. To me he was someone that was always embracing the latest and greatest. I presume, that may have become some of his (business) undoing. But his machine were gems.

                      I actually replaced the lasers on a Shelton with linear glass scales (I now remember the beam splitters again), changed the controller and added software to it that had vol comp in it, so no more mechanical sqaureness adjust. We took off some DEC computer, I think it was a PDP10 or MiniVAX if memory serves me here. The replacement was a DOS 286 clone.


                      Ah, great stories, Jan.
                      ***************************
                      PC-DMIS/NC 2010MR3; 15 December 2010; running on 18 machine tools.
                      Romer Infinite; PC-DMIS 2010 MR3; 15 December 2010.

                      Comment

                      Related Topics

                      Collapse

                      Working...
                      X